Agentic AI summits and conferences 2026 and 2027

A year ago, an agent was a slide in a keynote. Now there are summits with the word in the title, tracks inside every large AI event, and a vendor on every stand who says their product is one. That makes choosing an event harder, not easier, because the label has spread much faster than the experience behind it.

This page lists the events focused on agentic AI, drawn from the Wysera AI events directory, plus the broader AI conferences worth checking for an agent track. If you want the concept first, start with what agentic AI is, then come back.

An agent is software that is given a goal rather than a single question, and works towards it by deciding steps, using tools and acting on systems. The interesting problems start exactly where a chatbot stops: what is it allowed to touch, how do you know it did the right thing, and what happens when it does not.

So the test for an agentic AI event is whether the agenda spends real time on those problems. Sessions on evaluation (how teams check an agent's work at scale), permissions (what it can do without asking), failure handling (what it does when a tool returns nonsense) and oversight (where a person approves) are the signal. A programme of keynotes about the future of work, however good, is a general AI event wearing a new badge.

These are the upcoming events in the directory with agentic AI as their stated focus. The list updates from the directory and drops events once they finish. Each name links to its listing and on to the organiser's site, which is the only source to trust for dates, venue and tickets.

The group is wider than it looks. The Agentic AI Summit at UC Berkeley leans towards research and builders. The Agentic AI for Leaders Summit in London is pitched at executives deciding what to adopt. And finance has its own strand, with summits in New York on agentic AI in finance and banking, which we also cover alongside the rest of that sector in AI in finance conferences.

Most of the conversation about agents happens inside large general AI events rather than at dedicated summits. The ones below are broad business or engineering conferences. Agents are likely to come up at all of them, but how deeply varies by year and by track, so check the published agenda for agent sessions before you treat any of them as an agentic AI event.

The trade-off is predictable. A large event gives you range, a busy expo floor and the chance to see several vendors in a day. A dedicated summit gives you depth and a room full of people working on the same problem. If you are still deciding whether agents matter to you, go broad. If you have already decided and want to do it well, go narrow.

Format matters too. Recorded keynotes are a poor reason to travel, since most large events publish them afterwards. What you cannot get from a recording is the twenty minutes after a session with the person who built the system on stage, or a quiet comparison with someone at a company your size who tried the same vendor. If the trip is for those conversations, plan them before you go. If it is for the talks, wait for the videos and spend the money on a pilot.

Builders. Engineers and technical founders should favour research-adjacent and engineering events, where people show architectures, evaluation harnesses and what broke. The value is in comparing notes with someone who has shipped an agent and is willing to say what they would undo.

Leaders deciding what to adopt. Heads of operations, sales or marketing need a different room: one with practitioners from businesses that look like theirs, talking about cost, staffing and oversight. A leader-focused summit, or the business tracks of a larger event, will serve better than a technical programme.

Regulated industries. Finance, healthcare and legal teams should look for sector events, because their real questions are about audit trails, accountability and approvals, and those are the questions general events rush past.

Small-business operators. Honestly, most should not travel for this. The problems you will face adopting agents are practical: which workflow to start with, where a person signs off, how to check the output. Those are answered faster by a narrow pilot than by a summit. If you do go, pick an applied event where the other attendees have your budget and your constraints; our list of AI conferences for marketing and sales leaders is built for that reader.

An agentic AI event is only as useful as the questions you ask in the corridor. These four separate people who have run agents in production from people who have watched demos.

The first question matters most for a small team. The design choice we make at Wysera is that Wyse drafts and a person approves before anything sends, because the cost of an agent writing the wrong thing to a customer is higher than the cost of a click. You may draw the line elsewhere. What matters is that you draw it on purpose, and a summit is a good place to hear where others drew theirs.

The budget conversation is easier if you promise a decision rather than learning. Write one page before you go: the workflow you are considering handing to an agent, the two or three approaches you want to compare, the people or vendors at the event who can help you compare them, and the full cost of the trip. Approval for a specific bet comes much faster than approval for staying current.

While you are there, write one line about each useful conversation before the next one starts. Afterwards, send follow-ups within a few days and write the summary within a week, with a recommendation at the top: pilot this, wait on that, ignore the other. If the recommendation is to pilot, keep it narrow, give it a clear approval step and a date to review it, and judge it on the work it did rather than the demo you saw.

The rest of the AI calendar, by region and field, is in the AI events directory.